基于条件反射机制的用户行为习惯获取方法  

Method of User's Behavior Obtaining Based on Conditioned Reflex Mechanism

作　　者：路飞[1] 刘志勇[1] 田国会[1] 李健[1] 

机构地区：[1]山东大学控制科学与工程学院,济南250061

出　　处：《系统仿真学报》2014年第6期1274-1279,1290,共7页Journal of System Simulation

基　　金：国家自然科学基金项目(61305122;61075092);山东省自然科学基金(ZR2011FM011);山东大学自主创新基金项目(2012TS077;2011JC017);山东省博士后创新基金(201203059)项目资助

摘　　要：利用生物体的条件反射原理,提出了一种用户行为习惯获取方法。首先建立了统计聚类算法与基于低维聚类、高维验证的高维子空间聚类算法(SCLCHV,subspace clustering based on low dimensional clustering and high dimensional verification),对收集到的设备日常使用信息进行数据处理,获得用户的准行为习惯;在此基础上,设计习惯信息准确度评价函数来实现行为习惯信息的强化与淘汰,对获取的准行为习惯信息进行验证,从而获得准确的用户行为习惯信息。最后,对用户行为习惯获取策略进行仿真测试,测试结果表明该方法具有较高的准确性和较好的可行性。Imitated the organisms’ conditioned reflex principle, a new method for obtaining user’s behavior was proposed. Firstly, statistical clustering algorithm and an algorithm named subspace clustering based on low dimensional clustering and high dimensional verification were constructed to finish the extraction of the possible habits information through the collected daily using information to these equipments. After this, a behavior information accuracy evaluation function was designed to realize the reinforcement and elimination of the information. Then, the effective behavior information could be obtained after the verification and the test to the acquired possible habits information. At last, the simulation test to the method was done, and the simulation results show this method is effective.

关 键 词：用户行为习惯 资源管理 条件反射 聚类分析 准确度评价函数 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]